\begin{table}[ht]
\centering
\resizebox{.85\textwidth}{!}{
\begin{tabular}{lrrrr}
  \hline
Outcome & ITT & ITT\_D & CACE & p-value \\ 
  \hline
\multicolumn{5}{c}{I. Monitoring vs. Control}  \\ 
  \underline{A. White vs. Black} &  &  &  &  \\ 
  Index measure of favorable in-person interactions  & -0.015 & 0.81 & -0.018 & 0.388 \\ 
  Received post-visit callback  & -0.002 & 0.81 & -0.002 & 0.486 \\ 
  Received post-visit offer for unit  & -0.003 & 0.81 & -0.004 & 0.464 \\ 
  \underline{B. White vs. Hispanic} &  &  &  &  \\ 
  Index measure of favorable in-person interactions  & -0.061 & 0.81 & -0.075 & 0.14 \\ 
  Received post-visit callback  & -0.036 & 0.81 & -0.045 & 0.202 \\ 
  Received post-visit offer for unit  & -0.017 & 0.81 & -0.021 & 0.31 \\ 
  \underline{C. Black vs. Hispanic} &  &  &  &  \\ 
  Index measure of favorable in-person interactions  & -0.089 & 0.81 & -0.106 & 0.045 \\ 
  Received post-visit callback  & -0.035 & 0.81 & -0.043 & 0.206 \\ 
  Received post-visit offer for unit  & -0.014 & 0.81 & -0.017 & 0.328 \\ 
  \multicolumn{5}{c}{II. Punitive vs. Control (Upper Bound)}  \\ 
  \underline{A. White vs. Black} &  &  &  &  \\ 
  Index measure of favorable in-person interactions  & 0.007 & 0.718 & 0.01 & 0.555 \\ 
  Received post-visit callback  & 0.019 & 0.718 & 0.027 & 0.675 \\ 
  Received post-visit offer for unit  & 0.018 & 0.718 & 0.025 & 0.696 \\ 
  \underline{B. White vs. Hispanic} &  &  &  &  \\ 
  Index measure of favorable in-person interactions  & 0.048 & 0.718 & 0.067 & 0.806 \\ 
  Received post-visit callback  & -0.066 & 0.718 & -0.092 & 0.056 \\ 
  Received post-visit offer for unit  & -0.021 & 0.718 & -0.029 & 0.269 \\ 
  \underline{C. Black vs. Hispanic} &  &  &  &  \\ 
  Index measure of favorable in-person interactions  & 0.039 & 0.718 & 0.053 & 0.785 \\ 
  Received post-visit callback  & -0.085 & 0.718 & -0.119 & 0.02 \\ 
  Received post-visit offer for unit  & -0.039 & 0.718 & -0.055 & 0.123 \\ 
  \multicolumn{5}{c}{III. Punitive vs. Control (Lower Bound)}  \\ 
  \underline{A. White vs. Black} &  &  &  &  \\ 
  Index measure of favorable in-person interactions  & 0.007 & 0.8 & 0.009 & 0.555 \\ 
  Received post-visit callback  & 0.019 & 0.8 & 0.024 & 0.676 \\ 
  Received post-visit offer for unit  & 0.018 & 0.8 & 0.023 & 0.696 \\ 
  \underline{B. White vs. Hispanic} &  &  &  &  \\ 
  Index measure of favorable in-person interactions  & 0.048 & 0.8 & 0.061 & 0.806 \\ 
  Received post-visit callback  & -0.066 & 0.8 & -0.083 & 0.056 \\ 
  Received post-visit offer for unit  & -0.021 & 0.8 & -0.026 & 0.269 \\ 
  \underline{C. Black vs. Hispanic} &  &  &  &  \\ 
  Index measure of favorable in-person interactions  & 0.039 & 0.8 & 0.047 & 0.785 \\ 
  Received post-visit callback  & -0.085 & 0.8 & -0.107 & 0.019 \\ 
  Received post-visit offer for unit  & -0.039 & 0.8 & -0.049 & 0.122 \\ 
   \hline
\end{tabular}
}
\caption[Estimated Complier Average Causal Effects of Messages on Net Discrimination Levels for Monitoring-Control and Punitive-Control Comparisons]{\footnotesize Estimated Complier Average Causal Effects of Messages on Net Discrimination Levels for Monitoring-Control and Punitive-Control Comparisons. Cells contain estimates of the $ITT$, $ITT_D$ (proportion of Compliers), $CACE$, and $p$-values. $ITT$ and $ITT_D$ estimates are from OLS models with inverse probability weights and block fixed effects. $CACE$ estimates are from IV regression models with inverse probability weights and block fixed effects. Estimated $p$-values are reported in parentheses; $p$-values correspond to a one-sided test of the null hypothesis of equality of means for the monitoring-control and punitive-control comparisons, and to a two-sided test of the null hypothesis of equality of means for all analyses involving net discrimination against Hispanic (vs. black) testers. For the punitive-control comparison, the upper bound on the $CACE$ treats all subjects assigned to the punitive condition but who received the monitoring condition as subjects not complying with their treatment assignment. The lower bound on the CACE interprets the receipt of a monitoring message when assigned to punitive as effectively receiving a punitive message.}
\label{cace1}
\end{table}
